R Notebook

library(tidyverse)

dataset <- read_table("planes/data/Plane_TRAIN.txt", col_names = FALSE)

colnames(dataset) <- c("category", 1:145)

dataset <- dataset %>% add_column(alpha=1, .after = 1)

dim(dataset)

[1] 105 146

# get a description of the dataset

dataset %>% 
    group_by(category) %>%
    summarise(
        n = n()
    )

plot_dataset <- function(s_dataset) {
p_data <- s_dataset %>%
    rownames_to_column() %>%
    pivot_longer(-c(rowname, category, alpha)) %>%
    mutate(name = as.integer(name), category = as.factor(category))

    library(ggplot2)
    g <- ggplot(p_data %>% group_by(category), aes(x = name, y = value, color = category, group = rowname, alpha=alpha)) +
        geom_line() +
        theme_light() +
        theme(
            panel.border = element_blank(),
            axis.ticks.x = element_blank(),
            axis.ticks.y = element_blank(),
            panel.grid.major.x = element_blank(),
            panel.grid.minor.x = element_blank(),
            axis.title.x = element_text(size = rel(1), margin = margin(t = 20, r = 0, b = 0, l = 0)),
            axis.title.y = element_text(size = rel(1), margin = margin(t = 0, r = 20, b = 0, l = 0)),
        ) +
        labs(
            shape = "",
            alpha = ""
        ) +
        scale_alpha(guide = 'none') +
        xlab("time") +
        ylab("value")

    return(g)
}

plot_dataset(dataset)
ggsave("planes/results/planes.pdf", width = 10, height = 6)

source("methods/methods.R")

plot_outliers <- function(data, outliers) {
    data <- data %>% rownames_to_column() %>%
        mutate(alpha = if_else(rowname %in% outliers, 1, 0.1)) %>%
        mutate(category = if_else(rowname %in% outliers, "outlier", "no outlier")) %>%
        dplyr::select(-rowname)
    
    plot_dataset(data) +
        scale_color_manual(values=c("grey", "steelblue"))    
}

generate_outliers_dataset <- function(data, outliers, method) {
    data <- data %>% rownames_to_column() %>%
        mutate(alpha = if_else(rowname %in% outliers, 1, 0.1)) %>%
        mutate(category = if_else(rowname %in% outliers, "outlier", "no outlier")) %>%
        add_column(method = method) %>%
        dplyr::select(-rowname)
    data
}

data <- as.matrix(dataset[-c(1:2)])

aaknn <- AAKNN(data)
d1 <- generate_outliers_dataset(dataset, aaknn, "FAA + KNN")
plot_outliers(dataset, aaknn)

# ggsave("planes/results/FAAKNN.png")

aaknn <- OTRESH(data)
d2 <- generate_outliers_dataset(dataset, aaknn, "FOADATH")
plot_outliers(dataset, aaknn)

# ggsave("planes/results/FOADATH.png")

aaknn <- OMULTI(data)
d3 <- generate_outliers_dataset(dataset, aaknn, "FOADARMAH")
plot_outliers(dataset, aaknn)

# ggsave("planes/results/FOADARMAH.png")

aaknn <- RMAH(data, t = 1:144)
d4 <- generate_outliers_dataset(dataset, aaknn, "RMAH")
plot_outliers(dataset, aaknn)

# ggsave("planes/results/RMAH.png")

ise <- ISE(data, t = 1:144)
d5 <- generate_outliers_dataset(dataset, ise, "ISFE")
plot_outliers(dataset, ise)

# ggsave("planes/results/ISFE.png")

ogh <- OGH(data, t = 1:144)
d6 <- generate_outliers_dataset(dataset, ogh, "OUG")
plot_outliers(dataset, ise)

# ggsave("planes/results/OUG.png")

pond <- FADALARA(data, 2)
d7 <- generate_outliers_dataset(dataset, pond, "CROFADALARA")
plot_outliers(dataset, pond)

# ggsave("planes/results/CROFADALARA.png")

fom <- FOM(data)
d8 <- generate_outliers_dataset(dataset, fom, "FOM")
plot_outliers(dataset, fom)

# ggsave("planes/results/FOM.png")

fb <- FB(data)
d9 <- generate_outliers_dataset(dataset, fb, "FB")
plot_outliers(dataset, fb)

# ggsave("planes/results/FB.png")

trimknn <- TRIMKMEANS(data, k=7, out_rate=0.05)
d10 <- generate_outliers_dataset(dataset, trimknn, "TRIMKMEANS")
plot_outliers(dataset, trimknn)

# ggsave("planes/results/TRIMKMEANS.png")

trim <- TRIM(data, t = 1:144, out_rate=0.05)
d11 <- generate_outliers_dataset(dataset, trim, "TRIM")
plot_outliers(dataset, trim)

pond <- POND(data, t = 1:144, out_rate = 0.05)
d12 <- generate_outliers_dataset(dataset, pond, "POND")
plot_outliers(dataset, pond)

hdr <- HDR(data, t = 1:144, out_rate = 0.05)
d13 <- generate_outliers_dataset(dataset, hdr, "HDR")
plot_outliers(dataset, hdr)

d <- bind_rows(d1, d2, d3, d4, d5, d6, d7, d8, d9, d10, d11, d12, d13)

p_data <- d %>%
    rownames_to_column() %>%
    pivot_longer(-c(rowname, category, alpha, method)) %>%
    mutate(name = as.integer(name), category = as.factor(category))

g <- ggplot(p_data %>% group_by(category), aes(x = name, y = value, color = category, group = rowname, alpha=alpha)) +
    geom_line() +
    theme_light() +
    theme(
        panel.border = element_blank(),
        axis.ticks.x = element_blank(),
        axis.ticks.y = element_blank(),
        panel.grid.major.x = element_blank(),
        panel.grid.minor.x = element_blank(),
        axis.title.x = element_text(size = rel(1), margin = margin(t = 20, r = 0, b = 0, l = 0)),
        axis.title.y = element_text(size = rel(1), margin = margin(t = 0, r = 20, b = 0, l = 0)),
    ) +
    labs(
        shape = "",
        alpha = ""
    ) +
    scale_alpha(guide = 'none') +
    xlab("time") +
    ylab("value") +
    facet_wrap(~method) +
    scale_color_manual(values=c("grey", "steelblue"))    

g


ggsave("planes/results/outliers.pdf", width = 14, height = 8)